3
$\begingroup$

Let $(X,\|\cdot\|)$ be a Banach space with a Schauder basis and fix $p\in[1,\infty]$. Suppose that $X$ is asymptotic-$\ell_{p}$ with respect to this basis. It is known that the closed linear span of every (nontrivial) spreading model of $X$ is isomorphic to $\ell_{p}$ if $X$ is reflexive and at least contains an isomorphic copy of $\ell_{p}$ in general (replace $\ell_{p}$ by $c_{0}$ if $p=\infty$). In other words, the global asymptotic geometry of $X$ gives some information about the local asymptotic geometry.

Do there exist any known converse results? For example, are there general hypotheses that, in combination with the closed linear span of every spreading model containing an isomorphic copy of $\ell_{p}$, ensure that $X$ itself will be asymptotic-$\ell_{p}$?

Note that this question is cross-listed here on MSE.

$\endgroup$
2
  • 1
    $\begingroup$ It's good to see questions about asymptotic $\ell_p$ spaces. There is a lot of interesting literature that might help you. You should start by looking at the introduction to this paper: arxiv.org/abs/1902.10092 and references. I don't completely understand you statement: "It is known that the closed linear span of every (nontrivial) spreading model of 𝑋 is isomorphic to ℓ𝑝 if 𝑋 is reflexive and at least contains an isomorphic copy of ℓ𝑝 in general." Can you clarify? $\endgroup$ Commented Sep 9, 2020 at 11:22
  • $\begingroup$ Thanks for your comment and reference suggestion! To clarify my statement, it is Corollary 3.4.6 and the following paragraph of this thesis. $\endgroup$
    – JWP_HTX
    Commented Sep 9, 2020 at 14:08

1 Answer 1

4
$\begingroup$

The answer to the question you formulated is no in a very strong sense. For all $1<p<\infty$ there exists a reflexive space $X$ with an unconditional basis so that $X$ for all $\varepsilon>0$ every normalized weakly null sequence in $X$ admits a subsequence $1+\varepsilon$-equivalent to the unit vector basis of $\ell_p$ (so all spreading models are isomorphic to $\ell_p$) and yet its asymptotic structure contains $\ell_q^n$'s for some $p\neq q$. See Example 4.2 of this paper of Odell and Schlumprecht

There are many variation of the problem relating different asymptotic structures, and often the right question is to ask for a subspace with a better asymptotic structure under the assumption of, say, all spreading models being isomorphic to some $\ell_p$. The remarkable paper of Argyros and Motakis (that Kevin already referred to) gives some definite answers to some of these difficult questions. See the references therein to discover older results.

$\endgroup$
3
  • $\begingroup$ Thanks for your helpful answer! As it happens, I am interested in Banach spaces that are known to admit $\ell_{1}$ as a unique spreading model and that (I believe) cannot contain $\ell_{p}^{n}$'s. $\endgroup$
    – JWP_HTX
    Commented Sep 14, 2020 at 5:18
  • $\begingroup$ You need to be more specific. It is trivial to give counterexample to this. Take $\ell_1$-sum of $\ell^n_p$'s, $\left(\sum_n \ell^n_p\right)_{\ell_1}$ $\endgroup$ Commented Sep 14, 2020 at 16:30
  • $\begingroup$ You also want to ask copies of $\ell^n_p$'s as blocks in the space, otherwise you can always find them in $\ell_1$ for $p\le 2$. $\endgroup$ Commented Sep 14, 2020 at 16:38

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .